My data looks like this below:

> head(links_dat)
  set_num alllinks
1    None        8
2    None        3
3    None        9
4    None        3
5    None        2
6    None        2

There are four categories of set_num, alllinks is numeric. I want to plot the violin plot of the alllins across set_num, and also add the average value for corresponding distributions. I find the average of the third category is higher than the forth, but it shows a lower position in the plot. Can someone explain this for me?

Here is my plot code:

p <- ggplot(links_dat, aes(x = set_num, y = alllinks, fill = set_num)) +
  geom_violin(trim = FALSE) +   
  stat_summary(fun.y = mean, geom = "point", size = 0.5, color = "red") +
  scale_y_log10(labels = label_number()) +
  scale_fill_brewer(palette = "Blues") +
  theme_classic()